On some Phenomena in Vacuum-tubes. 



243 



The length of the electrode acted in the same -way as did the deflector 

 rod in Experiment 4. It is, therefore, clear that in arranging 

 Experiment 4 it was necessary to have the length of the rod some 

 multiple 6f the distance between the holes in the glass sphere and its 

 electrode ; also that the length of the glass rod should be some 

 definite multiple of the distance between the contraction of the tube 

 at the farther end and its electrode ; also that the length of the 

 electrodes should be short compared with any of these distances ; 

 otherwise a confusion of bands would be set up. 



The experiment with the disk of glass placed at the end of the 

 deflector rod showed that any effect which could be produced by this 

 disk alone was overpowered by the effects produced by the rod. 

 This would appear evident, since the two bands which would be 

 formed by the disk would practically be driven (or repelled) into one 

 by the action of the rod ; the independent action of the electrodes 

 in the case of Experiment 4 being absent may be due to the same 

 reason. 



I had considerable difficulty in getting over these points, for, at 

 first, confused results were presented in some tubes and not in others, 

 and it was only after investigating the matter, as here mentioned, 

 that I was able to construct tubes without difficulty to give definite 

 results when these various details were attended to. 



Experiment 7. 



Various forms of curved tubes are tried, some of them bent almost 

 into a circle. In several instances the little hollow glass spheres with 

 short rods bent to the curve of the tube are inserted, and in other 



R 2 



